Compare PDF Files Online – See Every Change Between Two PDFs

Compare PDF files online, without installing anything

Jobsixo gives you a quick way to compare PDF files online. Add the original document and the revised one, press Compare, and the tool lists every difference it finds: lines that were added, lines that were removed, lines that were reworded, pages that exist in only one file, and pages that look different even though their text did not change.

Both documents stay on screen side by side, so you are never matching a list of changes against a page you cannot see. Step through the differences one at a time and the page jumps to where each one sits. It is built for the everyday job of working out what someone changed between two versions of the same file.

How to compare two PDF files

1. 1
   Choose the original PDF in the first box and the revised PDF in the second. Each file can be up to 100 MB.
2. 2
   If either file is a scan, pick the OCR language beside the Compare button: English, Hindi, or Hindi + English. Files that already contain selectable text do not need it.
3. 3
   Press Compare. The text, the page images and the page counts are all checked on your own device.
4. 4
   Read the status line, then use Previous and Next difference to move through the list. The Highlight changes box switches the coloured marks on or off, and Synchronised scrolling, which starts switched on, keeps both pages level.
5. 5
   Use the page arrows, the Zoom slider and the Both, Original and Revised buttons to look more closely.
6. 6
   Press Report (text) or Report (JSON) if you need to send the list of changes to someone.

What the PDF comparison checks

A single check is never enough to say two documents are the same, so this PDF comparison tool looks at three things.

• Words. The text on each page is read line by line. A line that only exists in one file is reported as added or removed, and a line that mostly survived is reported as changed, showing the old wording and the new.
• Appearance. Each pair of pages is also compared as an image. If a page looks different but its text is identical, you are told, because that usually means a logo, a table or the layout moved.
• Page count. A page that exists in only one of the two files is reported as added or removed.

Pages are matched by position, so page 4 of the original is compared with page 4 of the revision. If someone inserted a page near the start, expect the pages after it to be flagged too. If you were looking for a PDF difference checker for finished documents rather than code, this is what it does.

How the text comparison decides what changed

For each page, the tool reads the text and groups it into lines by its position on the page. Every line on a page is then checked against the lines on the same page of the other file. A line with no exact match is a candidate difference. Candidates from the two sides are paired off when at least half of their words are shared, and a pair becomes one Changed entry that shows the old and new wording. Anything left unpaired is reported as Removed or Added.

That is why a corrected figure reads as a single change, while a deleted sentence reads as a removal. The half-the-words rule is a judgement call, so a heavily rewritten line may appear as one removal and one addition instead of a change. Both tell you the same thing.

How the visual comparison works

Text alone cannot tell you that a logo was swapped, a table was redrawn or a paragraph moved. So both versions of every shared page are also drawn as images, scaled to the same size, and compared pixel by pixel. Tiny differences from anti-aliasing are ignored, and a page is flagged only when enough of it has changed to matter.

Changed regions are marked on the Revised page when Highlight changes is on. A page appears in the list as a Visual change only if its text was identical. If the words differ too, the marks still show where the page changed, but the text entries lead the list.

A worked example

Suppose a landlord sends a revised rental agreement. The original has 3 pages and the revised file has 4. You add both and press Compare. The list shows four entries. On page 1, Monthly rent: Rs 18,000 has become Monthly rent: Rs 20,000, reported as Changed. On page 2, Notice period: 30 days has become 60 days, also Changed. On page 3 a new line, Security deposit: two months, is reported as Added. Page 4 exists only in the revised file, so it appears as Page added.

The text report summarises this as two modified lines, one added line and one added page, and it records the page count as 3 to 4. Nothing else changed, which is exactly what you needed to know before signing. Now imagine the landlord had inserted a new cover page at the front. Every page after it would have shifted by one and been flagged, because pages are matched by position. In that case, remove the extra page first, then compare.

Compare scanned PDFs with OCR

A scanned PDF is a set of pictures with no text underneath, so there is nothing to read until the words are recognised. When a file has no selectable text, the tool reads its pages with OCR before comparing, in the language you choose: English, Hindi, or Hindi + English.

OCR is not perfect. It can misread a character, so a difference reported on a scanned page may be a misreading rather than a real edit. Check the marked spot on the page before you rely on it. The first scanned comparison also needs to fetch the OCR engine and its language data, so it takes longer than comparing two typed documents.

How to read the results

The status line gives the total, for example four differences found, or states that none were found. Each row in the list carries a label (Added, Removed, Changed, Page added, Page removed or Visual change), the page number and a short description. Click a row to jump to that page.

The text report records when it was generated, both file names and page counts, a summary of counts, whether the words came from the files' own text or from OCR, and the full list. It ends with a Confidence and limits note. The JSON report holds the same data for anyone who wants to process it further.

Treat a result of no differences with some care. It means the words and the rendered pages matched within the tolerances above. It does not cover metadata, bookmarks or anything that never appears on the page.

Private by design

Your PDFs are opened and compared inside your own browser tab. The comparison does not upload them to a Jobsixo server, and nothing about the job needs an account. The one thing that is downloaded is the OCR engine with its language data, and only when you compare a scanned file.

When to compare two PDF files

• Contract revisions: confirm that the version you are about to sign only changed the clauses you agreed to change.
• Resume versions: see what you or a reviewer altered between drafts.
• Exam papers and question banks: check a reprint against the original before it goes out.
• Reports and proposals: find the edits made after the last review round.
• Forms and policies: spot the quiet change in an updated version of a document you already have.
• Quotations and price lists: catch a changed rate or a dropped item before the figures reach a client.

Tips for a cleaner comparison

• Compare like with like. Two exports of the same source, made with the same settings, produce the fewest false alarms.
• Put the older file in the Original box. Added and Removed are judged from that direction.
• Compare a section rather than a whole book when only part of it changed. Cutting the PDF down first makes the list shorter and the pages easier to match.
• Use the Original and Revised buttons to view one file at a time on a small screen.
• Download the text report before you close the tab. Nothing is stored for you afterwards.

Prepare your files first

A protected PDF has to be opened before it can be compared. If you hold the password, unlock the file first. A file that opens freely but blocks printing is also rejected here, because it is encrypted internally, and removing the restrictions fixes that. If a PDF is reported as damaged, repair the file and try again.

If the two versions arrive in pieces, merge the parts back together, and split a long file to cut it down to the section you need to check. A photographed document can be turned into a PDF with scan to PDF before comparing. Before you share a comparison report outside your team, redact private details from the documents it quotes.

Mistakes to avoid

• Swapping the two boxes and then reading Added as Removed.
• Comparing a scan with a typed copy and trusting every OCR difference.
• Ignoring the Visual change entries. They are how a swapped image or a moved table shows up.
• Comparing documents where a page was inserted early, then wondering why everything is flagged.

Limits and honest limitations

Lines are compared as sets on each page. If a sentence moves to a different position on the same page, the text comparison sees nothing, though the visual check may flag the page. If a line is repeated, two copies on one side and one on the other are not distinguished. A sentence that moves to another page shows as removed from one and added to the other.

Lines are built from text sitting at the same height, so on a page with two columns, text from both columns on the same line is joined into one long line. A change in either column then shows as a change to that long line.

OCR runs only when a file has almost no selectable text overall. A typed report with scanned pages at the back is read as typed, and the scanned pages contribute no words. The visual comparison still catches changes on them. When pages differ in size, the revised page is drawn at the original's size first, so tiny layout differences between paper sizes are smoothed out.